The_Signal_And_The_Nate.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "The-Signal-And-The-Nate.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyO9JwcRzQihFYs52Apf5uSR",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/microprediction/microprediction/blob/master/The_Signal_And_The_Nate.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IG3JjYDNIFQ7"
      },
      "source": [
        "# FiveThirtyEight versus PredictIt\n",
        "Ex-post Brier and likelihood for FiveThirtyEight presidential state results in the 2020 Election. \n",
        "\n",
        "See article at www.microprediction.com/blog\n",
        "\n",
        "We perform both a naive comparison and one with a Longshot adjustment. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z_OaoEaEwB09"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WonAhiCZwCx7"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qJAhpMQgIdLJ"
      },
      "source": [
        "### Nate Silver's pre-election probabilities"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rO-SUYBIV8BV"
      },
      "source": [
        "import pandas as pd\n",
        "import math\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "nate = pd.read_csv('https://projects.fivethirtyeight.com/2020-general-data/presidential_state_toplines_2020.csv')\n",
        "latest = nate.loc[nate['modeldate']=='11/3/2020',:]\n",
        "trump = latest[['state','winstate_inc']]\n",
        "nate = dict(zip(trump['state'].values,trump['winstate_inc'].values))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PDYVrj2dIaAj"
      },
      "source": [
        "### PredictIt pre-election state prices\n",
        "Swing states only at first ... see further down for more states. Tediously copied from PredictIt's election map. Let me know if I got anything wrong! "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zyzfEqG82LA_"
      },
      "source": [
        "predictit_swing = {'Arizona':0.43,'Iowa':0.63,'Georgia':0.57,'Florida':0.58,'Minnesota':0.25,\n",
        "            'New Hampshire':0.21,'Pennsylvania':0.38,'North Carolina':0.47,'Texas':0.71,\n",
        "            'Michigan':0.30,'Wisconsin':0.28,'Ohio':0.68,'Alaska':0.88}\n",
        "\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8Vrz_b8FJaoj"
      },
      "source": [
        "### Results\n",
        "Right now we leave some states very close to 1 or 0 if they aren't called. I used Betfair prices for this. Results aren't super-sensitive to those approximations."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z4kkT3b-JdHa"
      },
      "source": [
        "trump_swing = {'Arizona':0,'Iowa':1,'Georgia':0.04,'Florida':1,'Minnesota':0,\n",
        "            'New Hampshire':0,'Pennsylvania':0,'North Carolina':0.94,'Texas':1,\n",
        "            'Michigan':0,'Wisconsin':0,'Ohio':1,'Alaska':1}"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "clQlnN7dI_bF"
      },
      "source": [
        "### Compute Brier scores\n",
        "i.e. RMSE  of pre-post probability"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z3YiUh5p54ZN",
        "outputId": "0c01c329-8cce-48cd-a473-e4e5f4863d11",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "nate_swing = dict([(s,p) for s,p in nate.items() if s in predictit_swing])\n",
        "nate_brier = [ (p-trump_swing[s])**2 for s,p in nate_swing.items() ]\n",
        "predictit_brier = [ (p-trump_swing[s])**2 for s,p in predictit_swing.items() ]\n",
        "math.sqrt(np.mean(nate_brier)),math.sqrt(np.mean(predictit_brier))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.35474296511428793, 0.3530417277581526)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3iEnxHzzKDGM"
      },
      "source": [
        "Pretty close!  It is very hard for any one model to compete with a prediction market so in that sense, congrats to the FiveThirtyEight team.   \n",
        "\n",
        "## Longshot Effect Adjustment\n",
        "If you are not familiar with the LongShot effect, read the article \"The Signal and the Nate\" at www.microprediction.com/blog\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yDZq-OTEAh7k",
        "outputId": "ed5ee50f-5d98-4998-f566-750fe9c73587",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        }
      },
      "source": [
        "def adjusted_brier(expon:float,predictions:dict,result:dict)->float:\n",
        "   \"\"\" Compute Brier score after longshot adjustment  \"\"\"\n",
        "   adjusted = dict([(s,p**expon/(p**expon+(1-p)**expon)) for s,p in predictions.items()])\n",
        "   return math.sqrt(np.mean([ (p-result[s])**2 for s,p in adjusted.items() ]))\n",
        "\n",
        "# We choose exponents that might be vaguely realistic economically\n",
        "# If I had to guess, 1.3 is probably not far off reality for betting markets\n",
        "# and as it happens, that would have served FiveThirtyEight well also.  \n",
        "# But the plot below shows that any choice > 1 leads to the conclusion that \n",
        "# FiveThirtyEight lost out to PredictIt this time around. \n",
        "\n",
        "expons =  np.linspace(0.75,2,100)\n",
        "pi_briers = [ adjusted_brier(expon,predictit_swing,trump_swing) for expon in expons]\n",
        "nate_briers = [ adjusted_brier(expon,nate_swing,trump_swing) for expon in expons]\n",
        "plt.plot(expons,nate_briers,expons,pi_briers)\n",
        "plt.grid()\n",
        "plt.legend(['538','PredictIt'])\n",
        "plt.xlabel('Longshot effect exponent')\n",
        "plt.ylabel('Brier score')\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'Brier score')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Kc1p7rF8LXCt"
      },
      "source": [
        "This isn't a good picture for FiveThirtyEight. Let's see if likelihood serves them any better. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8Mxied8GJ7mo",
        "outputId": "3e68fb95-34c2-4b5a-ebe4-d120aedacbbb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        }
      },
      "source": [
        "import math\n",
        "def adjusted_loglike(expon,predictions,result):\n",
        "   fixed = dict([(s,p**expon/(p**expon+(1-p)**expon)) for s,p in predictions.items()])\n",
        "   return -np.mean([ math.log(p) if result[s] else math.log(1-p) for s,p in fixed.items() ])\n",
        "expons =  np.linspace(0.5,2,100)\n",
        "pi_ll = [ adjusted_loglike(expon,predictit_swing,trump_swing) for expon in expons]\n",
        "nate_ll = [ adjusted_loglike(expon,nate_swing,trump_swing) for expon in expons]\n",
        "plt.plot(expons,nate_ll,expons,pi_ll)\n",
        "plt.grid()\n",
        "plt.legend(['538','PredictIt'])\n",
        "plt.xlabel('Longshot effect exponent')\n",
        "plt.ylabel('Negative Log-Likelihood')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'Negative Log-Likelihood')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vKpCt7TiLv8h"
      },
      "source": [
        "No such luck. If anything the picture here is worse for FiveThirtyEight, even if you make no adjustment for LongShot. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4HsOdoNuiIf2",
        "outputId": "ce804840-44e3-4c7c-cf4f-a3b17f3b938e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "adjusted_loglike(1,predictit_swing,trump_swing),adjusted_loglike(1,nate_swing,trump_swing)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.4175868026706624, 0.4345285761879868)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DEMb3ZHhL-g7"
      },
      "source": [
        "### So how does FiveThirtyEight lose out to PredictIt? "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c75pS5r79x3C",
        "outputId": "621164b5-63c9-42d7-e0bb-1b4d87842134",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "comparison = [ (s,t,predictit_swing[s],nate_swing[s]) for s,t in trump_swing.items()]\n",
        "comparison"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[('Arizona', 0, 0.43, 0.319775),\n",
              " ('Iowa', 1, 0.63, 0.6013),\n",
              " ('Georgia', 0.04, 0.57, 0.418475),\n",
              " ('Florida', 1, 0.58, 0.309),\n",
              " ('Minnesota', 0, 0.25, 0.0419),\n",
              " ('New Hampshire', 0, 0.21, 0.10915),\n",
              " ('Pennsylvania', 0, 0.38, 0.1569),\n",
              " ('North Carolina', 0.94, 0.47, 0.360675),\n",
              " ('Texas', 1, 0.71, 0.617525),\n",
              " ('Michigan', 0, 0.3, 0.053175),\n",
              " ('Wisconsin', 0, 0.28, 0.056425),\n",
              " ('Ohio', 1, 0.68, 0.546725),\n",
              " ('Alaska', 1, 0.88, 0.8505)]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7kgReZ8ZMGr1"
      },
      "source": [
        "So we see that against the naive interpretation of PredictIt, FiveThirtyEight picks up Brier in Arizona, Georgia, Minnessota, New Hampshire, Pennsylvania, Michigan, Wisconsin.  Loses big in Florida. Loses in Iowa, North Carolina, Texas, Ohio and Alaska. \n",
        "\n",
        "On the other hand..."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kfl1gLdGM64f",
        "outputId": "0961fe30-caa6-4b5d-e42b-923954d2e501",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "expon = 1.3\n",
        "predictit_swing_adj = dict([(s,p**expon/(p**expon+(1-p)**expon)) for s,p in predictit_swing.items()])\n",
        "comparison_adj = [ (s,t,predictit_swing_adj[s],nate_swing[s]) for s,t in trump_swing.items()]\n",
        "comparison_adj"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[('Arizona', 0, 0.40940962293076516, 0.319775),\n",
              " ('Iowa', 1, 0.6663854245742022, 0.6013),\n",
              " ('Georgia', 0.04, 0.5905903770692347, 0.418475),\n",
              " ('Florida', 1, 0.6033888240750176, 0.309),\n",
              " ('Minnesota', 0, 0.19337992781764063, 0.0419),\n",
              " ('New Hampshire', 0, 0.15156189591295302, 0.10915),\n",
              " ('Pennsylvania', 0, 0.34605790462003205, 0.1569),\n",
              " ('North Carolina', 0.94, 0.4610322831808, 0.360675),\n",
              " ('Texas', 1, 0.7620586417812818, 0.617525),\n",
              " ('Michigan', 0, 0.2494613381806368, 0.053175),\n",
              " ('Wisconsin', 0, 0.22656669309192073, 0.056425),\n",
              " ('Ohio', 1, 0.7270890370763118, 0.546725),\n",
              " ('Alaska', 1, 0.9302255517774538, 0.8505)]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cGQeZyzlNpno"
      },
      "source": [
        "We see that the unwinding the longshot effect increases the discrepancy in states like Texas and Ohio, Alaska. FiveThirtyEight can't recover from that. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jCTHNfq9PD9f"
      },
      "source": [
        "## Adding more states \n",
        "\n",
        "I find it problematic that FiveThirtyEight does worse in swing states. But let's add some more states in and see if it makes a difference. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A7O604G2MoAH"
      },
      "source": [
        "predictit_more = {'California':0.06,'Washington':0.06,'Oregon':0.097,\n",
        "                  'North Dakota':0.96,'South Dakota':0.97,'Nebraska':0.97,\n",
        "                  'Kansas':0.96,'Virginia':0.07,'Mississippi':0.94,\n",
        "                  'Louisiana':0.97,'Alabama':0.97,'West Virginia':0.96,\n",
        "                  'Maryland':0.05}\n",
        "\n",
        "trump_more = {'California':0,'Washington':0,'Oregon':0,\n",
        "              'North Dakota':1,'South Dakota':1,'Nebraska':1,'Kansas':1,\n",
        "              'Virginia':0,'Mississippi':1,'Louisiana':1,'Alabama':1,\n",
        "              'West Virginia':1,'Maryland':0}\n",
        "predictit_more.update(predictit_swing)\n",
        "trump_more.update(trump_swing)\n",
        "nate_more = dict([(s,p) for s,p in nate.items() if s in predictit_more])\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pZ7tIzMCN2P-",
        "outputId": "5198fd47-68f6-49b2-d19e-eceb43d582e6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        }
      },
      "source": [
        "# Does it make a difference??\n",
        "# Short answer is no, and including more states makes it harder to improve the 538 model. \n",
        "\n",
        "expons =  np.linspace(0.5,2,100)\n",
        "pi_briers = [ adjusted_brier(expon,predictit_more,trump_more) for expon in expons]\n",
        "nate_briers = [ adjusted_brier(expon,nate_more,trump_more) for expon in expons]\n",
        "import matplotlib.pyplot as plt\n",
        "plt.plot(expons,nate_briers,expons,pi_briers)\n",
        "plt.grid()\n",
        "plt.legend(['538','PredictIt'])\n",
        "plt.xlabel('Longshot effect exponent (more states)')\n",
        "plt.ylabel('Brier score')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'Brier score')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vU8DHlAoxNkp"
      },
      "source": [
        "\"not really\". Add more states as you see fit, but be mindful of the 5% charge for withdrawals and 10% charge for winnings. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L7ERswBRwEmG",
        "outputId": "89eb2454-219d-4221-df85-db80dca2bb50",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 551
        }
      },
      "source": [
        "# Quick look at some alt data but here the prob model isn't calibrated\n",
        "cpolls = [0.591,0.540,0.498,0.586,0.372,0.467,0.434,0.443,0.529,0.504,0.365,0.605,0.432,0.554,0.521,0.562,0.597,0.559,0.474,0.387,0.376,0.486,0.480,0.552,0.573,0.561,0.579,0.476,0.484,0.433,0.455,0.401,0.499,0.619,0.539,0.622,0.444,0.518,0.427,0.522,0.597,0.583,0.525,0.552,0.386,0.465,0.443,0.642,0.473,0.659]\n",
        "a = 2\n",
        "cprobs = [ math.exp(a*p)/(math.exp(a*(1-p))+math.exp(a*p)) for p in cpolls ]\n",
        "states = [\"Alabama\",\"Alaska\",\"Arizona\",\"Arkansas\",\"California\",\"Colorado\",\"Connecticut\",\"Delaware\",\"Florida\",\"Georgia\",\"Hawaii\",\"Idaho\",\"Illinois\",\"Indiana\",\"Iowa\",\"Kansas\",\"Kentucky\",\"Louisiana\",\"Maine\",\"Maryland\",\"Massachusetts\",\"Michigan\",\"Minnesota\",\"Mississippi\",\"Missouri\",\"Montana\",\"Nebraska\",\"Nevada\",\"New Hampshire\",\"New Jersey\",\"New Mexico\",\"New York\",\"North Carolina\",\"North Dakota\",\"Ohio\",\"Oklahoma\",\"Oregon\",\"Pennsylvania\",\"Rhode Island\",\"South Carolina\",\"South Dakota\",\"Tennessee\",\"Texas\",\"Utah\",\"Vermont\",\"Virginia\",\"Washington\",\"West Virginia\",\"Wisconsin\",\"Wyoming\"]\n",
        "cascend = dict(zip(states,probs))\n",
        "cascend_more = dict([(k,v) for k,v in cascend.items() if k in trump_more])\n",
        "cascend_briers = [ adjusted_brier(expon,cascend_more,trump_more) for expon in expons]\n",
        "nate_briers = [ adjusted_brier(expon,nate_more,trump_more) for expon in expons]\n",
        "sum(cascend_briers),sum(nate_briers)"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-1-794a27d46e1e>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0mcpolls\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;36m0.591\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.540\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.498\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.586\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.372\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.467\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.434\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.443\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.529\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.504\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.365\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.605\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.432\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.554\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.521\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.562\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.597\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.559\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.474\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.387\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.376\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.486\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.480\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.552\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.573\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.561\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.579\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.476\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.484\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.433\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.455\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.401\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.499\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.619\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.539\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.622\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.444\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.518\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.427\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.522\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.597\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.583\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.525\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.552\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.386\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.465\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.443\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.642\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.473\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.659\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0ma\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0mcprobs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m \u001b[0mmath\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmath\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mmath\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mp\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcpolls\u001b[0m \u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m \u001b[0mstates\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m\"Alabama\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Alaska\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Arizona\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Arkansas\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"California\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Colorado\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Connecticut\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Delaware\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Florida\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Georgia\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Hawaii\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Idaho\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Illinois\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Indiana\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Iowa\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Kansas\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Kentucky\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Louisiana\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Maine\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Maryland\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Massachusetts\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Michigan\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Minnesota\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Mississippi\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Missouri\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Montana\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Nebraska\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Nevada\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"New Hampshire\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"New Jersey\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"New Mexico\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"New York\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"North Carolina\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"North Dakota\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Ohio\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Oklahoma\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Oregon\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Pennsylvania\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Rhode Island\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"South Carolina\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"South Dakota\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Tennessee\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Texas\"\u001b[0m\u001b[0;34m...\n\u001b[1;32m      6\u001b[0m \u001b[0mcascend\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstates\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mprobs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-1-794a27d46e1e>\u001b[0m in \u001b[0;36m<listcomp>\u001b[0;34m(.0)\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0mcpolls\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;36m0.591\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.540\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.498\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.586\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.372\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.467\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.434\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.443\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.529\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.504\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.365\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.605\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.432\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.554\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.521\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.562\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.597\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.559\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.474\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.387\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.376\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.486\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.480\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.552\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.573\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.561\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.579\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.476\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.484\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.433\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.455\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.401\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.499\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.619\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.539\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.622\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.444\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.518\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.427\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.522\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.597\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.583\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.525\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.552\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.386\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.465\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.443\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.642\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.473\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m0.659\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0ma\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0mcprobs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m \u001b[0mmath\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmath\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0mmath\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mp\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcpolls\u001b[0m \u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m \u001b[0mstates\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m\"Alabama\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Alaska\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Arizona\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Arkansas\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"California\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Colorado\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Connecticut\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Delaware\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Florida\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Georgia\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Hawaii\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Idaho\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Illinois\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Indiana\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Iowa\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Kansas\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Kentucky\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Louisiana\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Maine\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Maryland\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Massachusetts\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Michigan\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Minnesota\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Mississippi\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Missouri\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Montana\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Nebraska\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Nevada\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"New Hampshire\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"New Jersey\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"New Mexico\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"New York\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"North Carolina\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"North Dakota\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Ohio\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Oklahoma\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Oregon\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Pennsylvania\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Rhode Island\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"South Carolina\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"South Dakota\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Tennessee\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\"Texas\"\u001b[0m\u001b[0;34m...\n\u001b[1;32m      6\u001b[0m \u001b[0mcascend\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstates\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mprobs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'math' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ACm_-DNizNXp",
        "outputId": "9df16353-bccf-4a04-e536-2f7280094379",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "len(cascend_briers), len(nate_briers)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(100, 100)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0e1lOEp0zNwi",
        "outputId": "f905e2ff-1b36-4829-90c8-0aff75235d19",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 297
        }
      },
      "source": [
        "\n",
        "expons =  np.linspace(0.75,15,100)\n",
        "cascend_briers = [ adjusted_brier(expon,cascend_more,trump_more) for expon in expons]\n",
        "nate_briers = [ adjusted_brier(expon,nate_more,trump_more) for expon in expons]\n",
        "plt.plot(expons,nate_briers,expons,pi_briers)\n",
        "plt.grid()\n",
        "plt.legend(['FiveThirtyEight','Cascend'])\n",
        "plt.xlabel('Longshot effect exponent')\n",
        "plt.ylabel('Brier score')\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'Brier score')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}